Visualization

Martin Morgan
October 29, 2014

Three approaches

• 'base' graphics
• lattice

• ggplot2

  • A neat ggplot2 example: slide 19+ from Vince Carey's presentation.

Genome-centric

• Gviz for visualizing genomic regions
• ggbio provides extensive options

Interactive visualization

• rtracklayer for visualizing genomic ranges and managing UCSC genome browser sessions.
• epivizr for interactive display
• shiny web for interactive apps.

Reports

• ReportingTools for easy report templates

And…

• RColorBrewer and web site: helps choose sensible color schemes.
• Accidental aRt
• stick figure

Lab

Gviz / ggbio

Work through vignette section 2 the Gviz package, and then the 'Plotting fold changes in genomic space' in the RNASeq lab.

Peruse the vignette of the ggbio package.

Run the following DESeq2 work flow to arrive at a top-table; coerce the result to a data.frame.

library(DESeq2)
library(airway)
data(airway)
se = airway
dds <- DESeqDataSet(se, design = ~ cell + dex)
dds$dex <- relevel(dds$dex, "untrt")
dds <- DESeq(dds)
res <- results(dds)
resdf <- as.data.frame(res)

A 'volcano plot' shows the relationship between P-value and log fold change. Here's a basic volcano plot using base graphics; create a volcano plot using ggplot2 and / or lattice.

plot(-log10(padj) ~ log2FoldChange, resdf)